Hidden sample size change may have altered result
The IRIS trial in this months British Journal of Obstetrics & Gynaecology (BJOG) compared two speeds for administering MgS04 to mothers for fetal neuroprotection. The idea was that the slower one might reduce side effects. The trial report is here. It was registered here. The planned sample size was 140, and the primary outcome any maternal adverse effect. The authors recruited 51 women and found no difference. But they hid their failure to achieve the intended sample by writing, at best incorrectly: “We aimed to recruit 51 women in total.”
Fortunately the editor asked Dwight Rouse, one of the most experienced obstetric triallists in the world, to write a commentary. He checked the trial registration site and picked up the sample size change: “the authors failed to acknowledge it, but should have, the intended sample size when they originally registered the trial was 140 women”.
Good for him. Not good for the editors to fail to make the authors correct the manuscript. Online readers will likely miss Rouse’s commentary.
But it gets worse. The trial registration site now reads “Prospectively registered, target sample size 51, status complete”. Only on the trial registration history here will you find that on 3 March 2013, the sample size was altered from 140 to 51 because “Sample size was based on a reduction in the experience of any adverse effects of the magnesium sulphate infusion (the primary outcome).” This makes no sense. If the rate of the primary endpoint was lower than anticipated, it would be a reason for increasing the sample size!
Does it matter? The results were as follows:
Any adverse event. 15/25 (60%) 60-minute loading; 21/26 (81%) 20-minute loading; risk ratio (RR) 0.74; 95% confidence interval (95% CI) 0.51-1.08. Adverse event leading to ceasing the loading treatment (1/25 in the 60-minute loading group; 2/26 in the 20-minute loading group; RR 0.52; 95% CI 0.05-5.38.
The authors conclusion:
A slower rate of administering the loading dose of magnesium sulphate did not reduce the occurrence of maternal adverse effects overall.
But if the trial had gone on to its planned sample size of 140 and if the percentage of adverse events in each group among the latter 89 had been the same as the first 51 – supposition, but a reasonable supposition – the results might have been 42/70 (60%) v 56/70 (81%). RR 0.74, 95% CI 0.59 – 0.92. P = 0.0078 i.e. slow infusion does reduce side effects.
I know I’m a boring obsessive, but if the question matters, and this one surely does because MgSO4 treatment reduces cerebral palsy, then getting the right answer matters.
This was government funded research. If pharma did this sort of thing, imagine the fuss AllTrials.net would make.
BB Sophia opens
Sweden is far ahead of the UK when it comes to using independent maternity health providers.
Last week a new independent maternity hospital, BB Sophia, (click here) opened in Stockholm. Not a midwife led unit for normal deliveries, or a “too posh to push” place for the super rich to have Caesareans, but a full service maternity hospital, with midwives, obstetricians, anaesthetists and and its own neonatal unit. It will deliver 4,000 babies a year, about the same as the average of the five government-run units in Stockholm. And it’s the second large independent one in the city (click here for the other). They both receive funding per birth from the government, so care remains free to patients.
Compare with the UK. Last year a private midwife unit opened in Cheshire, and the only other one in the country, Tooting in London, closed due to lack of support. And that’s it. The whole of the UK, with a population six times larger than Sweden, doesn’t have a single independent maternity hospital serving NHS patients.
But Sweden is like us in one respect. The usual band of establishment providers are complaining that the new unit will undermine the government ones. The BMA would feel right at home.
Female genital mutilation (FGM), the excision all or part of the foreskin/prepuce, clitoris or labia, which causes in some cases labial fusion and narrowing of the vaginal opening, is illegal in the UK.
Campaigners want people who do it, and parents who allow it, to be prosecuted, and last week the director of public prosecutions, Alison Saunders, announced the first one. It’s an odd story. Here is the DPP statement.
“It was alleged that following a patient giving birth in November 2012, a doctor at the Whittington Hospital, in London, repaired FGM that had previously been performed on the patient, allegedly carrying out FGM himself.
“Having carefully considered all the available evidence, I have determined there is sufficient evidence and it would be in the public interest to prosecute Dr Dhanoun Dharmasena for an offence contrary to S1 (1) of the Female Genital Mutilation Act (2003).
“I have also determined that Hasan Mohamed should face one charge of intentionally encouraging an offence of FGM, contrary to section 44(1) of the Serious Crime Act (2007), and a second charge of aiding, abetting, counselling or procuring Dr Dharmasena to commit an offence contrary to S1 (1) of the Female Genital Mutilation Act (2003).”
The newspapers report that Dr Dharmasena had been an obstetrician at the Whittington in 2012 and that Mr Mohamed was related to the alleged victim. What could possibly have happened?
The wording “repaired FGM that had previously been performed on the patient, allegedly carrying out FGM himself” suggests that a midline scar joining the labia across the anterior part of the vagina had torn during the delivery. This is a common post FGM injury. The torn edges bleed so obstetricians are obliged to repair them. They can either place sutures to stop the bleeding but leave the labia separate, i.e return the anatomy more nearly to normal, or they can suture the labia back together across the vaginal opening, i.e. return to the post FGM pre-delivery anatomy.
Although the patient or her partner may request the latter for cultural or sexual reason, it is generally advised against. The delivery has resolved one of the harmful effects of the FGM, and it could be argued that the doctor harms the woman a second time by rejoining the labia. But it’s a stretch to call such a repair FGM. Of course, if the doctor actually removed part or all of a healthy prepuce, clitoris, or labia, that would be FGM, but the DPP’s statement does not imply that.
I dealt with many such cases in the early 1980s when I worked in Africa. My practice was to compromise. I sutured the anterior part of the labia across the vagina but left the posterior part open. I argued that, by leaving a more open vagina, I was making intercourse easier for the woman, but that by repairing the anterior part, I was also respecting the couple’s culture.
It was a more innocent time. I’m glad there were no lawyers looking over my shoulder.
On the face of it this an extraordinary prosecution but perhaps there’s more to it. Ripe-tomato.org will watch the case with interest.
A long review of the short story writer in this week’s New Yorker (click here). An early story, The Thirteenth Woman, is two sentences and 138 words long. Not as short as Hemingway’s six word: “For sale: baby shoes, never worn”, but short. Here it is.
The Thirteenth Woman
In a town of twelve women there was a thirteenth. No one admitted she lived there, no mail came for her, no one spoke of her, no one asked after her, no one sold bread to her, no one bought anything from her, no one returned her glance, no one knocked on her door, the rain did not fall on her, the sun never shone on her, the day never dawned for her, the night never fell for her; for her the weeks did not pass, the years did not roll by; her house was unencumbered, her garden unattended, her path not trod upon, her bed not slept in, her food not eaten, her clothes not worn; and in spite of all this she continued to live in the town without resenting what it did to her.
Glans amputation at neonatal circumcision
Last year ripe-tomato.org drew readers attention to the complications mentioned in a recent WHO manual for teaching neonatal circumcision (click here). Some were serious and probably under-reported; clumsy doctors don’t publicise their mistakes. Parents may sue, but rarely publicise their son’s injury either, and he certainly won’t. We wondered if parents were fully informed.
A recent report (J Paed Urol, Dec 2013, click here, full text here glans amputation) should give cause for thought. Some Canadian urologists describe six cases of accidental amputation, referred over a five or ten* year period. Two were successfully reattached, albeit with deformity; it’s too soon to know whether sensation was affected. The other three required multiple reconstructive operations. Warning – it’s illustrated.
It’s a nice paper. The authors show clearly how the problem happens when the foreskin has not been fully separated before surgery, and how it can be avoided by dividing adhesions between the glans and foreskin. Doctors performing circumcision should read it carefully. Parents will want to ensure their son gets done by an expert – or perhaps leave him intact.
* The abstract says the injuries occurred over a five year period, but the main text says over ten. Simple error. Mistakes happen!
What might it look like
If tongue-tie division improves breast feeding, tens of thousands of babies a year might undergo it in the UK. Sooner or later things would go wrong – scissors would slip, and babies with clotting disorders would bleed heavily. That doesn’t mean we shouldn’t do it. Just that we should be sure it does more good than harm. What might a good quality trial look like?
Someone with skills in systematic reviewing and no vested interest, should work with the Cochrane Collaboration to pull together the previous trial evidence and decide whether the question has already been answered. Presumably this is in hand.
Funders also need evidence that the association between tongue tie and breast feeding difficulties is real: a study of babies feeding nicely and babies with problems, with tongues all examined by an expert, without knowledge of how the breast feeding is going. If tongue-tie is equally common in both groups we can stop worrying about it. Has there been such a study?
The trial must compare division with never dividing. A trial comparing immediate with delayed division has two problems. It precludes measuring any long term outcomes; even duration of breast feeding, let alone anything like speech development or long term health, and it limits participation to staff and parents who already believe it works. Measures of things like nipple pain and breast feeding scores are particularly susceptible to biased reporting, and may be reduced, at least in theory, because the baby is in pain from the division. No-one wants to treat breast pain by hurting the baby.
Would any parent agree to their baby joining such a trial? Many will want one or other option, and recruitment will be a challenge. But many trials have been successfully conducted in other similarly polarised fields. It’ll need many centres, and many parent invitations. No-one would be forced to join.
Some babies in the “no division” group might get divided – it would be impossible and unethical to forbid division forever. To avoid biasing the results in favour of “no division” such babies would be analysed in their original group, by “intention to treat”. But to keep their numbers small, only parents who can give a reasonable commitment to stick with the allocated treatment should join. This will also slow recruitment.
By the time independent randomisation, trial registration, data management, analysis by intention to treat, agreed sample sizes and defined endpoints that don’t get altered are included, the trial might cost £1M or so. But let’s say each division costs £100 and 10,000 babies a year are eligible. If no trial is done the NHS could spend that every year for ever!
If the results are negative the trial would have saved thousands of babies from unnecessary surgery, and the money saved can be spent on other ways to help breast feeding. If division improves breast feeding, it can be rolled out properly, with incalculable health benefits.
My obituary of her from AOdeadpool (click here)
The appeal of Shirley Temple, child star of such 1930’s Hollywood tripe as Bright Eyes, Curly Top and Heidi waned when she reached puberty. She forged a successful second career as politician and diplomat, married and buried two husbands, became US ambassador to Ghana and Czechoslovakia, and survived breast cancer by forty years. But she was lucky to escape Hollywood unscathed. Watch any You Tube clip, e.g The good Ship Lollipop (click here). What was going on?
The novelist Graham Greene got the subtext immediately, and in 1938 accused the studio bosses of “procuring Shirley Temple for immoral purposes” and “middle-aged men and clergymen” of getting off on her under-age sexuality. He was successfully sued for his troubles by 20th Century Fox, who managed to convince a judge that he was libelling the actress, rather than her minders. But he was right. The middle-aged men saw it then. We can all see it now. I
Shirley Temple’s youthful flirting
Made her popular with perverts
Now we find it too disturbing
Looking up her miniskirts.
To Graham Green the films were loathsome
But the judge refused his claim.
Thought they were entirely wholesome.
Now he’d hang his head in shame
But she escaped her prurient bosses.
Outlived two husbands, even cancer.
Made her career in political causes.
Topped it off as US ambassador.